Mathieu Malaterre

Results 170 comments of Mathieu Malaterre

would be nice to have this for 1.0.1. thanks

closing since fixed in 0a9ec963b8478c45af963e78eb65e332fb4d07d7

@mo271 if the tests are supposed to always fail, simply mark it so: ``` set_tests_properties(always_failing_test_name PROPERTIES WILL_FAIL TRUE) ```

@szabadka Here is what log reveal: ``` Get:162 https://deb.debian.org/debian unstable/main amd64 libjxl-testdata all 0.0~git20220711.d81acac-1 [24.9 MB] ``` Which seems to be latest git commit: * https://github.com/libjxl/testdata/commits/main Also: ``` $ dget...

ouch ! ``` Testing oss-fuzz/clusterfuzz-testcase-minimized-djxl_fuzzer-5914266367557632 ./lib/jxl/base/file_io.h:81: JXL_FAILURE: Failed to open file for reading: /usr/share/libjxl-testdata/oss-fuzz/clusterfuzz-testcase-minimized-djxl_fuzzer-5914266367557632 ./lib/jxl/testdata.h:22: JXL_CHECK: ReadFile(full_path, &data) ``` I did not pay attention that `SIGILL, Illegal instruction.` was a...

For reference: * https://stackoverflow.com/questions/6049563/with-fno-exceptions-what-happens-with-new-t

@jonsneyers Here is the difference: https://en.cppreference.com/w/cpp/memory/new/nothrow ``` #include #include int main() { try { while (true) { new int[100000000ul]; // throwing overload } } catch (const std::bad_alloc& e) { std::cout

@jonsneyers Re-reading your answer, I think I now understand your point. Something like this: > If libjxl fails to do a memory allocation at any point in the codebase, it...

@jonsneyers I conducted a small experiment on my Debian bullseye machine (default `libstdc++` installation). ``` % cat foo.cxx #include void foo() // call foo() from main() { while (true) {...